dbt Labs dbt-Analytics-Engineering dbt Analytics Engineering Certification Exam Exam Practice Test

Hook 1

on-run-end: "{{ macro_x() }}"

Answer: macro_1

Hook 2

models:

:

post-hook: "{{ macro_x() }}"

Answer: macro_2

Hook 3

on-run-start: "{{ macro_x() }}"

Answer: macro_3

Hook 4

{{

config(

pre-hook: "{{ macro_x() }}"

)

}}

Answer: macro_4

dbt supports run-level hooks and model-level hooks. Run-level hooks fire once per invocation, while model-level hooks fire around each individual model.

on-run-end is a run-level after hook that executes once after the entire dbt command completes. Because macro_1() must run after every dbt run, it correctly belongs here.

The post-hook configured under the models: section runs after each model in that scope finishes building. This matches the requirement for macro_2() to execute after a model runs.

on-run-start is a run-level before hook and fires once before dbt begins executing any models for that command, making it the right place for macro_3() which must run before every dbt run.

Finally, the pre-hook specified inside a model’s config() block runs before that specific model is built. Since macro_4() must execute before a model runs, it belongs in the pre-hook configuration.

Thus the correct mapping is:

on-run-end → macro_1

model post-hook → macro_2

on-run-start → macro_3

model pre-hook → macro_4.

The purpose of this question is to determine how to validate that the input values in the state column support the business logic that determines the is_active field. Since the logic checks whether state = 'Active', it is critical that the state column only contains values that the business process expects. In the example shown, acceptable states appear to be: 'active', 'churned', and 'trial'.

The correct way to enforce this expectation is to apply an accepted_values test on the state column. This ensures that any unexpected state (e.g., 'inactive', 'pending', 'deleted', or NULL) will cause the test to fail, alerting the team that the upstream system is producing unexpected or invalid values. Additionally, adding not_null ensures every user record contains a valid state.

Option A is the only configuration that applies the accepted values test to the correct column (state) and reflects the expected domain of values.

Options B and D incorrectly apply tests to the derived column is_active, not the source column that needs validation. Option C only checks nullability and uniqueness, which does not validate the range of allowed values and thus does not protect the business logic.

Therefore, Option A is the only correct answer.

In dbt, when you configure a generic test like not_null in YAML, the configuration for that test must be a mapping, not another list item. The correct structure is:

tests:

- not_null:

config:

where: "state = 'completed'"

In your original YAML, you wrote:

tests:

- not_null:

- config:

where: "state = 'completed'"

The extra dash (- config) makes config an element of a list rather than a key under the not_null test. This breaks the expected shape of the test definition. When dbt parses the schema.yml, it fails to correctly interpret the patch for the stg_tasks model, which leads to the warning: “Did not find matching node for patch with name 'stg_tasks'…”.

By removing the extra dash and nesting config directly under not_null, dbt now reads this as a single generic test named not_null with a config block that passes the where argument. This allows dbt to correctly attach the test to the completed_at column of the stg_tasks model and eliminates the compiler warning.

Therefore, Option B is the only structurally valid configuration and is the correct answer.

1. Can only be defined on SQL models

Answer: contract

2. Errors are returned before the model is built

Answer: contract

3. May be configured to customize severity

Answer: test

4. May be run on ephemeral models

Answer: test

dbt tests and contracts serve different purposes in ensuring data quality and model correctness. Tests evaluate data after it is produced, while contracts validate the structure of a model before dbt attempts to build it.

A model contract is schema-level enforcement that describes required columns, data types, and constraints. Contracts can only be applied to SQL models, not Python or ephemeral models. Because contracts validate the model’s schema before executing any SQL, dbt surfaces those errors before the model is built, preventing invalid schemas from being deployed.

In contrast, tests evaluate the data after dbt builds a model. Tests can be written generically (unique, not_null, relationships, accepted_values) or as custom SQL tests. They run after the model is materialized. Tests also allow severity configuration, enabling warnings instead of failures for less critical issues—something contracts do not support.

Tests also run on ephemeral models, because dbt expands ephemeral SQL inline within downstream models, allowing tests to still execute logically against the resulting compiled SQL. Contracts, however, do not apply because ephemeral models never materialize into database objects.

Thus:

“SQL-only” and “errors before build” → contract

“custom severity” and “run on ephemeral models” → test

If you want the next question formatted the same way, send it!

The correct answer is C: Underlying data must be accessible on your data platform.

dbt does not perform data ingestion or data loading. Instead, dbt operates after raw data is already available in your warehouse. This means that before dbt can build any models—whether staging, intermediate, or mart-layer models—the underlying source data must already exist and be accessible in the connected data platform (Snowflake, BigQuery, Redshift, Databricks, etc.). dbt uses SQL to transform existing relations; therefore, if the data platform cannot access the underlying tables or external sources, model execution will fail.

Option A is incorrect because sources do not need to be defined before building models. Models can be built without using sources at all. Source definitions are optional metadata and lineage declarations, not prerequisites.

Option B is incorrect because service accounts are not required; dbt can connect through any credential mechanism supported by the warehouse (OAuth, user accounts, tokens, etc.).

Option D is incorrect because dbt itself performs transformations on raw data—cleaning raw data beforehand is not required; in fact, that is one of dbt’s main responsibilities.

Thus, the only true prerequisite is that the warehouse must contain accessible underlying data.

This test is a singular test, meaning it is defined as a standalone SQL file placed inside the /tests directory (not attached through YAML like generic tests). dbt executes singular tests by running the SQL query, and the test fails if the query returns any rows. In this case, the test selects all records from fct_orders where status_code = 13. That means if even one row in fct_orders has a status code of 13, the test will fail, because dbt interprets returned rows as evidence of a data-quality issue.

The configuration block at the top sets severity='error', meaning the test raises an error when it fails, rather than a warning. This setting overrides default severity behavior, so option B is incorrect. Singular tests do not need to be attached in YAML because they run automatically when placed under /tests → making option A incorrect. Meanwhile, generic tests live in /tests/generic, but singular tests should live directly under /tests (option C incorrect).

Therefore, the only correct statement is that records with status_code = 13 will cause the test to fail, because the test query is explicitly written to identify such rows.

Check your terminal or log file for the output from the most recent dbt run.

Look for specific error messages associated with the models that failed.

3️⃣ Isolate the problem by dbt run --select model_name to run a single model and confirm whether the issue is localized to that model.

4️⃣ Use dbt run --select model_name+ to run the model and its downstream dependencies, ensuring that your fix works across the DAG.

Brief Explanation

First, you always inspect the latest run output (step 1).

Then, identify the exact failing models and error messages (step 2).

Next, you reproduce the issue on the individual model to be sure the fix works locally (step 3).

Finally, you re-run the model plus its downstream dependencies to validate the fix across the DAG (step 4).

Update int_order_items by replacing:

{{ source('tpch', 'orders') }} → {{ ref('stg_tpch_orders') }}

{{ source('tpch', 'lineitem') }} → {{ ref('stg_tpch_line_items') }}

In dbt’s recommended modeling pattern, sources should be referenced only in the staging layer. Downstream models (intermediate and mart/fact models) should reference staging models via ref(), not the raw source() directly. This keeps all raw-to-clean logic centralized in the staging layer and makes refactoring and documentation easier.

In the original DAG, int_order_items depends directly on tpch.orders and tpch.lineitem via source(), and staging models also depend on those same sources. That creates parallel paths from the sources and breaks the clean layered architecture.

By updating the int_order_items SQL to use:

from {{ ref('stg_tpch_orders') }}

join {{ ref('stg_tpch_line_items') }} ...

instead of:

from {{ source('tpch', 'orders') }}

join {{ source('tpch', 'lineitem') }} ...

you ensure that int_order_items sits entirely downstream of the staging layer. The new DAG becomes linear: sources → staging → intermediate (int_order_items) → any further marts. This improves modularity, reusability of cleaned logic, and makes the DAG easier to reason about and maintain.

dbt model contracts enforce that the column names, data types, and number of columns defined in the contract exactly match the columns produced by the compiled SQL. If any column appears in one location (the contract or the SQL) but not in the other, dbt raises an enforcement error.

In this scenario, the error message shows:

ORDER_ID is missing in the model definition, meaning the SQL no longer contains a column named order_id, but the contract still expects it.

ORDER_KEY is missing in the contract, meaning the model SQL now contains a new column, but this column has not been added to the contract.

To fix the mismatch, you must remove columns from the contract that no longer exist in the SQL and add to the contract any new columns that now appear in the SQL.

Therefore:

Option A — Remove order_id from the contract — is correct because the column no longer exists in the model SQL.

Option D — Add order_key to the contract — is correct because the SQL now produces this column.

Options B and C incorrectly alter the wrong side of the definition, and Option E would create a mismatch in the opposite direction.

Thus, the correct fixes are A and D.

dbt Python models support a limited set of materializations because they rely on execution within the data platform’s Python compute engine (such as Snowpark for Snowflake, Dataproc for BigQuery, or Spark). These engines require models to materialize into actual relations—tables or views—in order to persist the results of Python-based transformations.

The ephemeral materialization, however, is fundamentally incompatible with this behavior. Ephemeral models do not create relations in the warehouse; instead, dbt inlines their SQL logic directly into downstream models. Since Python models cannot be inlined (they execute Python code, not SQL), dbt does not allow ephemeral Python models. dbt requires Python model outputs to be materialized as either:

table

view

incremental

Therefore, ephemeral is not supported for Python models, and attempting to configure a Python model as ephemeral will result in a compilation error.

The reason is straightforward: ephemeral logic depends on SQL compilation, while Python models depend on executing Python code in the data platform. Because these mechanisms are incompatible, dbt restricts Python models to relational materializations only.

Thus, the correct answer is No — ephemeral is not supported for Python models.

Will always point to the latest version of the source schema

Correct Match: ✔ defer

Allows to use objects built in the target schema with any downstream tool

Correct Match: ✔ dbt clone

3️⃣ Allows to safely modify objects built in the target schema

Correct Match: ✔ defer

4️⃣ Is a point-in-time operation

Correct Match: ✔ dbt clone

defer and dbt clone serve very different purposes in dbt, and understanding their behavior is essential for choosing the right command.

The --defer flag tells dbt to use already-built objects from a previous environment (often production) instead of rebuilding them. Because it always references the existing target schema’s most recent objects, it “always points to the latest version of the source schema.” Since no objects are overwritten when using defer, it also “allows safely modifying objects built in the target schema”—your development environment uses production objects without altering them.

By contrast, dbt clone creates a point-in-time copy of existing relations. This cloned schema is static; it does not auto-update when source data changes. Therefore, clone is classified as a “point-in-time operation.” Since clone copies physical tables/views into a new schema, downstream tools (BI dashboards, ML pipelines) can safely query the cloned environment without affecting production, making “allows to use objects built in the target schema with any downstream tool” the correct match.

Thus, defer is used for logic substitution without copies, while clone is used for replicable, point-in-time snapshots.

This error is caused by invalid Jinja syntax, specifically by nesting {{ }} blocks inside another Jinja expression. The expression:

{{ source("{{ env_var('input') }}", 'table_name') }}

compiles to:

{{ source({{ env_var('input') }}, 'table_name') }}

Here, Jinja sees {{ inside another {{ ... }} block. Jinja does not allow nested print statements like this; instead, functions should be called directly inside a single pair of curly braces. The parser encounters an unexpected } where it expects part of a valid expression (hence “expected token ':', got '}'”), which is a classic symptom of mismatched or nested curly braces.

The correct usage is:

select * from {{ source(env_var('input'), 'table_name') }}

In this form, env_var('input') is evaluated first, and its result is passed as the first argument to source() within one Jinja expression.

Option C is therefore the correct debugging approach: inspect your Jinja and look for incorrectly nested curly brackets. Options A and D are generic and don’t address the root cause, while B talks about quoting in SQL, which is not the problem—the error arises before SQL compilation, at Jinja parse time.

Modularity in dbt means breaking transformations into small, reusable steps that are easy to reason about, test, and reuse. In this DAG, two staging models (stg_detail_categories and stg_details) feed into two mart-level models (skills_with_details and lessons_with_details). If there is repeated logic in those two mart models—such as common joins, filters, or derived fields—that logic is a good candidate to be extracted into an intermediate model.

Option E follows dbt best practices: inspect the SQL of lessons_with_details and skills_with_details and identify shared logic to move into a new intermediate model (for example, int_details_with_categories). Both downstream models can then reference this intermediate model using ref(), centralizing logic and improving maintainability.

Options A and B push joins/unions into the staging layer, which should focus on clean, 1:1 representations of sources—not combining multiple sources. Option C collapses distinct staging models into one, reducing clarity and breaking the source-to-staging contract. Option D creates an overly wide, less reusable final table instead of focusing on shared reusable logic.

Therefore, the action that truly increases modularity is Option E: creating an intermediate model based on common SQL patterns.

The correct answers are B and E.

The dbt retry command was introduced to help developers quickly re-run only what failed in the prior dbt invocation. According to dbt documentation, dbt retry examines metadata from the previous execution, including the run results and the manifest.json, to determine which nodes failed. This aligns with Option E, as dbt uses the manifest to identify the models, tests, and other nodes that errored or were not executed successfully.

Option B is also correct because dbt retry works only when the previous run failed due to execution errors (e.g., a warehouse failure or a logic error in SQL). If the failure was due to a syntax error that prevented dbt from compiling, there will be no valid manifest produced, so retry cannot execute. Thus, dbt retry can only re-run models when the previous failure occurred after successful compilation.

Options A and C are incorrect because retry does not run all nodes nor does it automatically pick up exactly where an execution stopped — instead, it runs failed and downstream unbuilt nodes as determined by the manifest. Option D is also incorrect; retry uses the execution metadata, not selectors, to determine what should run.

1. There is a config block at the top of the model with materialized = 'table'

➡️ Answer: False

2. The corresponding table already exists in the data warehouse

➡️ Answer: True

3. The incremental model was executed with the --full-refresh flag

➡️ Answer: False

4. The corresponding table does not already exist in the data warehouse

➡️ Answer: False

The is_incremental() macro in dbt is used within an incremental model to determine whether the current invocation should perform an incremental update or a full rebuild. It evaluates to True only during an incremental run when the target table already exists in the data warehouse and dbt is not performing a full refresh.

For is_incremental() to return True, two conditions must be met:

The model must be materialized as incremental, not table or view.Therefore, if the config specifies materialized = 'table', the model is not incremental, and the function evaluates to False.

The table already exists in the target schema.In this case, dbt will run the model in incremental mode, making is_incremental() return **True`.

If the incremental model is run with --full-refresh, dbt intentionally rebuilds the entire table, meaning is_incremental() evaluates to False, even if the table exists.

Lastly, if the target table does not exist yet, dbt must create it from scratch, so the run is treated as a full rebuild, not an incremental update, causing is_incremental() to evaluate to False.

Thus, the only scenario where is_incremental() returns True is when the table exists and the model is actually incremental.

Information Type

Retrieved From

Singular tests

C. .sql files

Column data types

A. Data platform information schema

Generic tests

B. .yml configuration

SQL code

C. .sql files

Column descriptions

B. .yml configuration

Model dependencies

C. .sql files

The dbt docs command compiles metadata about your project by gathering information from three primary sources: your warehouse’s information schema, your YAML configuration files, and your SQL model files. Understanding which metadata comes from which source is essential for debugging and for effective documentation practices.

Singular tests live inside .sql files within the /tests directory. Since dbt renders these tests directly from SQL files, their definitions appear in documentation sourced from that location.

Column data types come from the warehouse itself. dbt introspects the data platform information schema to retrieve actual types because dbt does not infer or define column types—only the warehouse does.

Generic tests (e.g., unique, not_null, accepted_values) are declared in .yml files. These YAML definitions contain test configurations, descriptions, and parameters, which dbt uses to document and execute these tests.

SQL code for models is naturally sourced from .sql files where the models are defined. This includes logic such as SELECT statements, CTEs, and transformations.

Column descriptions are written exclusively in .yml files. dbt never extracts descriptions from SQL comments—only from YAML.

Model dependencies come from the ref() and source() calls inside .sql model files, which dbt parses to build the DAG.

Operation

Correct Location

Aggregations (booked_hotel, booked_ride)

In a model dedicated to pre-aggregate data for reporting

Standardizing customer_id

In the first layer of models

Selecting customer_name

In the first layer of models

Joining employees

In a model between the first layer and final layer of models

In dbt’s recommended modeling framework, transformations should be centralized according to their purpose and level of abstraction. Basic cleaning and column standardization—such as shortening customers.id using left(customers.id, 12)—belongs in the first layer of models, commonly known as staging models. This layer is responsible for producing clean, consistent, analytics-ready fields. Selecting raw descriptive fields like customers.name also belongs in this layer.

Joins that enrich a dataset by combining cleaned staging outputs across domains—such as joining customers with employees—belong in intermediate models. These models unify business logic that sits between the raw staging layer and the final aggregations used for reporting.

Finally, aggregations like counting types of accommodations (booked_hotel and booked_ride) belong in marts or reporting models, especially when they involve summarization that downstream tools (dashboards, BI reports) will consume. This keeps heavy business logic centralized and reusable.

The only correct statement is Option C, because dbt allows you to override the schema for a package’s models directly within your project’s dbt_project.yml file. Packages behave just like subfolders within your own project, and therefore the standard configuration pattern applies:

models:

timeformat:

+schema: timeseries

This correctly overrides the package’s default schema, allowing all models within the timeformat package to be built into the timeseries schema.

Why the other options are incorrect

Option A

Local packages can be added using a path reference, but that is not what a private Git repository refers to. A private repo must be added using git: plus SSH/HTTPS URL, not local: unless the developer literally cloned the repo into a local directory. The question specifically describes a private repository, so this is not the correct statement.

Option B

Packages are installed using dbt deps, not dbt build.

dbt build only runs models, tests, and snapshots; it does not install dependencies.

Option D

Including a version or revision is allowed for private Git-based packages. You can specify branches, tags, or commit hashes safely in packages.yml.

In dbt, a singular test is any SQL file placed directly inside the tests/ directory—not inside the tests/generic/ folder. Singular tests contain a custom SQL query where returning any rows means failure. They are distinct from generic tests, which are YAML-defined and live inside the generic directory, often referencing macros.

Looking at the folder structure:

In tests/generic/

furniture_customers_test.sql → This is a generic test, not a singular test.

In the root tests/ directory

macro_stg_tpch_orders_assert_pos_price.sql → singular test

macro_stg_tpch_suppliers_assert_pos_acct_bal.sql → singular test

stg_tpch_orders_assert_positive_price.sql → singular test

When you run:

dbt test --select 'test_type:singular'

dbt selects only singular tests, meaning tests that are individually written SQL files in the top-level tests/ directory.

Therefore, dbt will run:

macro_stg_tpch_orders_assert_pos_price

macro_stg_tpch_suppliers_assert_pos_acct_bal

stg_tpch_orders_assert_positive_price

It will not run furniture_customers_test, because that file is inside tests/generic/, which is reserved for template-based generic tests and is not classified as singular.

Thus, the correct answer is Option C.